Color Spaces in OpenCV | Python

Color spaces are models used to represent and organize the color channels in an image. They define how colors are formed from primary colors and stored, allowing computers to process and manipulate color information efficiently. Each color space has a specific purpose and the choice of color space depends on the task. In this article, we will see more about color spaces in OpenCV, how they are used and their core concepts.

Common Color Spaces in OpenCV

OpenCV provides several color spaces that are used in image processing. Let’s see the most important ones:

1. RGB (Red, Green, Blue)

RGB color model represents colors using three primary colors: red, green and blue. Each pixel in an image is a combination of these three colors at different intensities, creating a wide range of possible colors.This is the standard color space used for displaying images on screens and working with digital images.

2. BGR (Blue, Green, Red)

OpenCV uses BGR instead of RGB as its default color space where the red and blue channels are swapped. While this difference may seem small, it’s important to remember when working with OpenCV to avoid color mismatches.

3. HSV (Hue, Saturation, Value)

HSV is another color space used in OpenCV. It’s based on the human perception of color and it splits color information into three components:

• Hue: Represents the color itself (e.g red, blue, green).
• Saturation: Shows the vibrancy of the color.
• Value: Defines the brightness of the color.

The hue value ranges from 0 to 179, while saturation and value range from 0 to 255.

It is useful for tasks like color segmentation and object detection. For example, it allows us to easily isolate specific colors (like red or blue) in an image.

4. CMYK (Cyan, Magenta, Yellow, Black)

CMYK color model is used for color printing. It’s a subtractive color model, means colors are created by subtracting light using different combinations of cyan, magenta, yellow and black inks. In this model, colors become darker as we add more ink. It is used in print and graphic design but it's not as frequently used in OpenCV for image processing.

5. Grayscale

A grayscale image contains only shades of gray, with each pixel representing a different intensity level from black to white. It is used when color information is not needed such as in edge detection or image thresholding tasks.

Converting Between Color Spaces in OpenCV

OpenCV makes it easy to convert between different color spaces. This is useful when we want to perform tasks like object detection based on color or extract features using specific color information.

Let's see how to perform common color space conversions using OpenCV’s cv2.cvtColor() function. Here we are using a random sample image which you can download from here.

import cv2
from google.colab.patches import cv2_imshow 

image = cv2.imread('/content/sample_parrot image jpg')

image_rgb = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)

image_hsv = cv2.cvtColor(image, cv2.COLOR_BGR2HSV)

image_gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)

image_lab = cv2.cvtColor(image, cv2.COLOR_BGR2LAB)

cv2_imshow(image)
cv2_imshow(image_rgb)
cv2_imshow(image_hsv)
cv2_imshow(image_gray)
cv2_imshow(image_lab)

Output:

Practical Example

Lets see some common practical examples for better understanding.

1. Detecting Red Color in an Image Using HSV

Here we will see how to use the HSV color space to detect and isolate a specific color, in this case, red within an image. The HSV color space is often preferred over the RGB color space for tasks like color detection because it separates color (hue) from intensity (value). This makes it easier to define specific color ranges and perform operations like color segmentation.

• mask = cv2.inRange(image_hsv, lower_red, upper_red): cv2.inRange() creates a mask that isolates the red areas of the image based on the defined range.
• result = cv2.bitwise_and(image, image, mask=mask): The mask is applied to the original image using cv2.bitwise_and() to keep only the red regions.

import cv2
import numpy as np
from google.colab.patches import cv2_imshow


image = cv2.imread('/content/sample_parrot image jpg')

image_hsv = cv2.cvtColor(image, cv2.COLOR_BGR2HSV)

lower_red = np.array([0, 120, 70])
upper_red = np.array([10, 255, 255])

mask = cv2.inRange(image_hsv, lower_red, upper_red)

result = cv2.bitwise_and(image, image, mask=mask)

cv2_imshow(result)

cv2.waitKey(0)
cv2.destroyAllWindows()

Output:

The output image will display only the areas of the image that contain the red color, with all other areas turned black. This technique can be useful for various applications such as object detection, color segmentation and tracking specific colors in images.

2. Visualizing the Different Color Channels of an RGB Image

Now we will break down an RGB image into its individual color channels; Red, Green and Blue and display each channel separately. This process is important for understanding how each color component contributes to the final image. Visualizing the individual channels can be useful in image processing tasks such as color-based segmentation, image enhancement and feature extraction.

• B, G, R = cv2.split(image): Separate the image into its individual Red, Green and Blue channels using cv2.split().

import cv2
from google.colab.patches import cv2_imshow 

image = cv2.imread('/content/sample_parrot image jpg')

B, G, R = cv2.split(image)
cv2_imshow(image)
cv2.waitKey(0)

cv2_imshow(B)
cv2.waitKey(0)

cv2_imshow(G)
cv2.waitKey(0)

cv2_imshow(R)
cv2.waitKey(0)

cv2.destroyAllWindows()

Output:

• Blue Channel (B): This channel highlights the amount of blue in the image. Areas with more blue will appear brighter, while other areas will be darker or gray.
• Green Channel (G): Similar to the blue channel, this displays the green intensity. It plays a important role in natural images, particularly landscapes.
• Red Channel (R): This channel displays red intensity which influences the warmth of the image, highlighting reds, oranges and skin tones.

Applications of Color Spaces in Image Processing

1. Object Detection and Tracking (HSV): HSV is ideal for detecting specific colors in an image as it separates color (hue) from intensity (value), making it useful to varying lighting conditions and easier for tracking colored objects.
2. Image Segmentation (HSV, Grayscale): Color spaces like HSV and Grayscale simplify segmentation tasks by isolating relevant color or intensity information, helping to distinguish foreground from background in images.
3. Color Correction and Enhancement (RGB): RGB color space is used for adjusting color intensity and correcting image colors, improving visual appeal or compensating for poor lighting conditions in images.
4. Facial Recognition (Grayscale): Grayscale images are used in facial recognition because they focus on intensity rather than color, making it easier to detect features like edges and shapes.